Social Software Misuse

• Synopsis
• Project Outline
• People
• Related Publications

"Social Software Misuse" terms cases of anti-social behavior in online communities. An example for this is the distribution of spam via the e-mail infrastructure which is done by a small percentage of all mail users. Nevertheless, according to a recent study, the amount of spam mails sent per day reaches 95% of all mails sent per day. This makes automatized countermeasures such as spam filtering technologies a necessity. A surprising fact, however, is that spam is one of the few misuses for which detection technologies are being developed. Apart from them there are many other unaccounted social software misuses which threaten online communities, for instance vandalism and edit wars in Wikipedia. Goal of this project is research and development of new technologies for the automatic detection of social software misuse.

A service on the Internet (esp. on the World Wide Web) is called a social software if its purpose is online communication between two or more users. A social software therefore gathers a community of users who meet frequently on the infrastructure of the service. We distinguish eleven types of social software which are depicted in the following table.

In all online communities there are some participants who show an anti-social behavior. They misuse the social software in diverse ways and with diverse intentions, however, their actions always harm the welfare of the community. We distinguish three categories of misuses: destructive misuses, profit seeking misuses, and counterproductive misuses. Destructive misuses are meant to harm, impede, or destroy someone or something and profit seeking misuse are meant to raise one's personal profit by illegal or unethical actions. Both of the former are conducted deliberately whereas this is not the case with counterproductive misuses: here, the sum of one's otherwise well-intentioned actions forms the misuse. The following table gives an overview of misuses we have documented so far.

Our current efforts are directed at developing new automatic detection approaches for vandalism and edit wars in Wikis. Especially the Wikipedia community will benefit from such solutions.

• Martin Potthast
• Benno Stein

Students: Robert Gerling, Dennis Hoppe